Make run-pass/env-home-dir.rs test more robust #47718
Conversation
|
(rust_highfive has picked a reviewer for you, use r? to override) |
src/test/run-pass/env-home-dir.rs
Outdated
| @@ -18,16 +18,18 @@ use std::path::PathBuf; | |||
|
|
|||
| #[cfg(unix)] | |||
| fn main() { | |||
| let oldhome = var("HOME"); | |||
| let oldhome = var("HOME").unwrap_or("/home".into()); | |||
| let newhome = format!("{}/new", oldhome); | |||
There was a problem hiding this comment.
Why did you change to now construct a relative path instead of some random string?
It seems like before we never even looked at the value we got back, is that right?
There was a problem hiding this comment.
I could have only changed the line that assume the home_dir returns Some, that is
- assert!(home_dir().is_some());
+ assert_ne!(home_dir(), Some(PathBuf::from("/home/MountainView")));but that could fail if getpwuid_r returns "/home/MountainView" (or other chosen string). Setting newhome to a relative path avoids this situation.
There was a problem hiding this comment.
You're saying that could fail if set_var had no effect, but it just happens that the directory name is /home/MountainView? I suppose that's true. It just seems a bit weird for the windows test (below) to be asymmetric.
There was a problem hiding this comment.
I think I don't quite understand what is being fixed here =)
We're still asserting that home_dir returns Some -- how was the test failing before?
There was a problem hiding this comment.
(Anyway, the patch seems fine, just trying to understand. =)
There was a problem hiding this comment.
You're saying that could fail if set_var had no effect, but it just happens that the directory name is /home/MountainView? I suppose that's true. It just seems a bit weird for the windows test (below) to be asymmetric.
I changed the unix case because that's was failing for me, but in the end, I just need that the test does not assume home_dir return Some.
We're still asserting that home_dir returns Some -- how was the test failing before?
The test was failing because home_dir was returning None. Now we are using assert_ne!, so it pass if home_dir returns None.
There was a problem hiding this comment.
Ah, I missed the assert_ne thing somehow. Can you add a comment on top saying something like:
/// When HOME is not set, some platforms return `None`, but others return `Some` with a default.
/// Just check that it is not the `newhome_path`.
Also, I think it'd be best if the unix + windows tests match up -- I'm ok with either your new version or the old one, but I guess it's easier to just revert those commits.
nikomatsakis
added
the
S-waiting-on-author
Remove the assumption that home_dir always returns Some This allows the test to be executed with [cross](https://github.com/japaric/cross).
malbarbo
force-pushed
the
env-home-dir
branch
from
15abe46 to
898fdcc
Compare
|
Updated. |
|
@bors r+ -- thanks! |
|
📌 Commit adeb0ae has been approved by |
|
@bors rollup |
Make run-pass/env-home-dir.rs test more robust Remove the assumption that home_dir always returns Some. This allows the test to be executed with [cross](https://github.com/japaric/cross).
Make run-pass/env-home-dir.rs test more robust Remove the assumption that home_dir always returns Some. This allows the test to be executed with [cross](https://github.com/japaric/cross).
Remove the assumption that home_dir always returns Some.
This allows the test to be executed with cross.